Support Care Cancer DOI 10.1007/s00520-013-2112-0
ORIGINAL ARTICLE
Metastatic epidural spinal cord compression among elderly patients with advanced prostate cancer Benjamin A. Spencer & Jin Joo Shim & Dawn L. Hershman & Brad E. Zacharia & Emerson A. Lim & Mitchell C. Benson & Alfred I. Neugut
Received: 25 March 2013 / Accepted: 25 December 2013 # Springer-Verlag Berlin Heidelberg 2014
Abstract Background A recent randomized trial demonstrated that for metastatic epidural spinal cord compression (MESCC), a complication of advanced prostate cancer, surgical decompression may be more effective than external beam radiation therapy (RT). We investigated predictors of MESCC, its treatment, and its impact on hospital length of stay for patients with advanced prostate cancer. Methods We used the SEER-Medicare database to identify patients >65 years with stage IV (n=14,800) prostate cancer. We used polytomous logistic regression to compare those with
The linked SEER-Medicare database was used in this study. The interpretation and reporting of these data are the sole responsibility of the authors. B. A. Spencer : M. C. Benson Department of Urology, College of Physicians and Surgeons, Columbia University, New York, NY, USA B. A. Spencer : J. J. Shim : D. L. Hershman : A. I. Neugut Department of Epidemiology, Mailman School of Public Health, Columbia University, New York, NY, USA D. L. Hershman : E. A. Lim : A. I. Neugut Department of Medicine, College of Physicians and Surgeons, Columbia University, New York, NY, USA B. E. Zacharia Department of Neurological Surgery, College of Physicians and Surgeons, Columbia University, New York, NY, USA B. A. Spencer : D. L. Hershman : E. A. Lim : M. C. Benson : A. I. Neugut Herbert Irving Comprehensive Cancer Center, College of Physicians and Surgeons, Columbia University, New York, NY, USA A. I. Neugut (*) Columbia University Medical Center, 722 West 168th Street, Room 725, New York, NY 10032, USA e-mail: [email protected]
and without MESCC and those hospitalized for treatment with surgical decompression and/or RT. Results MESCC developed in 711 (5 %) of patients, among whom 359 (50 %) received RT and 107 (15 %) underwent surgery±RT. Median survival was 10 months. MESCC was more likely among patients who were black (OR 1.75, 95 %CI 1.39–2.19 vs. white) and had high-grade tumors (OR 3.01, 95 %CI 1.14–7.94), and less likely in those younger; with prior hormonal therapy (OR 0.73, 95 %CI 0.62–0.86); or with osteoporosis (OR 0.63, 95 %CI 0.47–0.83). Older patients were less likely to undergo either RT or surgery, as were those with ≥1 comorbidity. Patients with high-grade tumors were more likely to undergo RT (OR 1.92, 95 %CI 1.25–2.96). Those who underwent RT or surgery spent an additional 11 and 29 days, respectively, hospitalized. Conclusions We found that black men with metastatic prostate cancer are more likely to develop MESCC than whites. RT was more commonly utilized for treatment than surgery, but the elderly and those with comorbidities were unlikely to receive either treatment. Keywords Prostate cancer . Metastatic epidural spinal cord compression . Palliative care . SEER-Medicare
Introduction Each year, about 4 % of men diagnosed with prostate cancer are diagnosed with distant or metastatic disease [1]. The most common site for metastatic spread in prostate cancer is bone [2]. Bone metastases can cause pain, fractures, and metastatic epidural spinal cord compression (MESCC) [3]. The resulting edema, venous congestion, and demyelination can lead to irreversible spinal cord infarction if not treated promptly. With life expectancy for patients with MESCC estimated at 4 months, the decision to relieve the compression must be
Support Care Cancer
made with realistic goals [4]. However, the potential improvement in functional status (pain, ambulation, and urinary continence) and quality of life can be substantial. Relief of spinal cord compression can be achieved either with direct surgical decompression, external beam radiation therapy (RT) or both. In 2005, a randomized trial by Patchell et al. demonstrated a benefit from circumferential decompression, as compared to posterior laminectomy, followed by postoperative RT in patients with metastatic cancer presenting with signs and symptoms of cord compression and a true displacement of the spinal cord by an epidural mass [5]. For subjects who were randomized to both surgery and RT, 84 % were able to walk after treatment as compared to 57 % in the RT alone arm (p=0.001) and this ambulation persisted for 122 days as compared to 13 days in the RT alone group (p=0.003). Though this study is the only randomized trial evaluating the use of RT versus surgery for MESCC, it has been criticized because of slow patient accrual, suggesting a recruitment bias. The objective of our study was to identify predictors of MESCC, determine the patterns of use for surgery and RT, and investigate whether these interventions decrease the number of days spent in the hospital. We were also interested in whether disparities due to race or access to care influence the development of MESCC and its subsequent treatment.
whose only reporting source was autopsy or death certificate, whose reason for entitlement was not age, or whose date of death differed between SEER and Medicare by more than 3 months were also excluded. Among the remaining 139,627 patients, we selected a cohort that was diagnosed with American Joint Cancer Committee (AJCC) stage IV prostate cancer and survived at least 90 days following diagnosis. We then identified patients who had at least one billing claim with a diagnosis of spinal cord compression (ICD-9 Code 336.9) no more than 30 days prior to the prostate cancer diagnosis. This would allow for the possibility that the cord compression was the presenting symptom for the prostate cancer. Demographic variables SEER-Medicare provides the age, race/ethnicity, marital status, and tumor grade as categorical variables. Age categories are ordinal with 5-year increments between categories. Race/ethnicity was described as white, black, Hispanic, and Other/Missing. Marital status at time of cancer diagnosis was categorized as Married, Single/Divorced/Separated/Widowed, or Unknown. Tumor grade was categorized as well, moderately, poorly, and undifferentiated. Socioeconomic status
Methods Data source We analyzed data from the Surveillance, Epidemiology, and End Results (SEER)—Medicare database. The SEER database includes information on cancer diagnoses, tumor histology, stage of disease, treatments, socioeconomic status at the census tract and zip code level, survival and demographic characteristics. The Medicare database provides information on Medicare A (inpatient) and B (outpatient) eligibility, reason for Medicare entitlement, and diagnoses. The physician and hospital claim files, as well as inpatient claim files, were used to search for the claims on diagnoses or surgical procedures. The SEER and Medicare databases were combined by linking these patients by their unique patient identification number. Cohort selection We identified individuals who were 65 years or older at the time of cancer diagnosis and received a pathologically confirmed primary diagnosis of prostate cancer (SEER Site Code 54) as their first cancer between January 1, 1991, and December 31, 2007. Patients who were enrolled in a non-Medicare health maintenance organization or not covered by Medicare Part A and B over the same period were excluded. Patients
We followed the guidelines by Du et al. [6] to create a socioeconomic status score by equally weighting income, education, and poverty level provided from the 2000 census tract data. Patients were assigned a rank score from 0 to 4, with 0 being the lowest score. Approximately 1.6 % (N=233) of the cohort lacked sufficient data in one or more categories and were excluded from the analyses. Assessment of comorbid disease Using the comorbid conditions identified by Charlson et al. [7], we searched for diagnostic codes of the ICD-9 Clinical Modification and Healthcare Common Procedure Coding System (HCPCS) for 18 medical conditions. Each medical condition was assigned a weight and subsequently used to calculate the final index score. In order to obtain the complete diagnosis and medical claims, the physician and outpatient claims, as well as hospital claims, were included in the comorbidity calculation, as described by Klabunde et al. [8]. Claims submitted from 13 months prior to 4 months following the date of cancer diagnosis were considered in the comorbidity index calculation. Treatment characteristics Receipt of radiation and surgery were extracted from the Medicare database by searching the HCPCS, Current
Support Care Cancer
Procedural Terminology (CPT) codes, ICD-9 Clinical Modification diagnostic codes and procedure codes from the national claims history files, outpatient statistical analysis files, and Medicare provider analysis and review files. We excluded radiation and surgery claims that were billed more than 15 days prior or 60 days following the date of diagnosis of spinal cord compression because patients could have received these treatments for medical conditions other than spinal cord compression. The CPT codes 63001, 63003, 63005, 63012, 63015–63017, 63020, 63030, 63035, 63040, 63042–63048, 63050–63051, 63055–63057, 63064, 63066, 63075–63076, 63078, 63081–63082, 63065–83088, 63091, 63101–63013, 63275–63278, 63280–63283, 63286–63287, 63290, and 63295 and ICD-9 CM codes 03.0, 03.09, 80.5, 80.51, 80.52, 80.59, 81.0 81.00 81.04, 81.05, and 81.09 corresponded to surgical decompression. The ICD-9 CM codes V58.0, V66.1, V67.1, 92.21–92.26, 92.29–92.33, and 92.39 and CPT/ HCPCS codes 77401–77409, 77411–77414, 77416–77421, 77427, 77431–77432, 77435, 77470, and 77499 were used to identify patients who received radiation therapy. Based on the claims, patients were assigned to one of three treatment groups—radiation alone, surgery alone or in conjunction with radiation (+/- RT), and neither. Receipt of androgen deprivation therapy (ADT) with a GnRH agonist (HCPCS codes J1950, J3315, J9202, J9217-J9219, J9225) was assigned for patients who had at least one claim for receipt of ADT prior to the MESCC. Statistical analysis We used the chi-square test to compare the distributions of those who developed spinal cord compression with those who did not (see Table 1), and multivariable logistic regression to identify predictors of spinal cord compression, adjusting for the measured clinical and demographic characteristics of the patients. Polytomous logistic regression models were used to test associations between patient demographics and treatment assignment. All statistical tests were two-sided with an alpha of 0.05. Statistical analyses were performed using SAS version 9.2 (Cary, NC). We obtained approval for this study from the Columbia University Medical Center Institutional Review Board.
Results We identified 14,800 men with stage IV prostate cancer who met our eligibility criteria. Table 1 shows their demographic and clinical characteristics. The majority of the stage IV patients were white (79 %), married (67 %), lived in an urban area (89 %), had high-grade tumors (53 %), had no comorbidities (52 %), had used ADT prior to the diagnosis of
Table 1 Socio-demographic and clinical characteristics of men >65 years with stage IV prostate cancer in SEER-Medicare, 1991–2007 Characteristic
Age 65–69 70–74 75–79 80–84 85+ Race White Black Hispanic Other Marital status Married
Entire sample
Spinal cord compression sample
(n=14,800) n (%) 3,065 (20.7 3,925 (26.5 3,210 (21.7 2,568 (17.4 2,032 (13.7
(n=711) n (%) 170 (23.9 %) 225 (31.7 %) 149 (21.0 %) 102 (14.4 %) 65 (9.1 %)
%) %) %) %) %)
11,745 (79.4 %) 1,854 (12.5 %) 273 (1.8 %) 928 (6.3 %)
488 (68.6 %) 143 (20.1 %) 14 (2.0 %) 66 (9.3 %)
9,870 (66.7 %)
458 (64.4 %)
Single/divorced/widowed Unknown Residence Urban Rural Socioeconomic status First quintile Second quintile Third quintile Fourth quintile Fifth quintile Missing Tumor grade Well/moderately differentiated Poorly differentiated Undifferentiated Unknown Comorbidities
4,220 (28.5 %) 710 (4.8 %)
235 (33.1 %) 18 (2.5 %)
13,135 (88.8 %) 1,665 (11.3 %)
654 (92.0 %) 57 (8.0 %)
2,610 (17.6 3,223 (21.8 3,325 (22.5 2,936 (19.8 2,706 (18.3
158 (22.2 122 (17.2 146 (20.5 158 (22.2 127 (17.9
4,152 (28.0 %) 7,591 (51.3 %) 242 (1.6 %) 2,815 (19.0 %)
153 (21.5 %) 351 (49.4 %) 15 (2.1 %) 192 (27.0 %)
None One ≥Two Prior ADT use No Yes Prior osteoporosis No Yes Treatment None Radiation Surgery
7,696 (52.0 %) 3,871 (26.2 %) 3,233 (21.8 %)
399 (56.1 %) 176 (24.8 %) 136 (19.1 %)
5,815 (39.3 %) 8,985 (60.7 %)
370 (52.0 %) 341 (48.0 %)
12,828 (86.7 %) 1,972 (13.3 %)
57 (8.0 %) 654 (92.0 %)
NA NA NA
245 (34 %) 359 (50 %) 107 (15 %)
%) %) %) %) %)
%) %) %) %) %)
Support Care Cancer
MESCC (61 %), and had no prior history of osteoporosis (87 %). Almost 5 % (n=711) developed MESCC, among whom 359 (50 %) underwent palliative radiation therapy and 107 (15 %) underwent surgical decompression±RT. In the multivariable analysis, we found that MESCC was more likely to occur among those who were younger; black (OR 1.75, 95%CI 1.39, 2.19 vs. white); had undifferentiated tumors (OR 3.01, 95%CI 1.14, 7.94 vs. well differentiated); were diagnosed earlier in the cohort; did not use ADT prior to the diagnosis of MESCC (OR 0.73, 95%CI 0.62, 0.86); and had no prior history of osteoporosis (OR 0.63, 95%CI 0.47, 0.83) (Table 2). There were 711 stage IV patients (4.8 %) who developed MESCC. Of these, 359 (50 %) underwent RT, while 107 (15 %) underwent spinal surgery. Among subjects who developed MESCC, older patients were less likely to undergo either radiation or surgery. Similarly, those with comorbidities were also less likely to undergo either radiation (2+ comorbidities: OR 0.53, 95%CI 0.34, 0.82 vs no comorbidities) or surgery± radiation (2+ comorbidities: OR 0.40, 95%CI 0.20, 0.78 vs no comorbidities). Patients with high-grade tumors were more likely to undergo RT (poorly/undifferentiated: OR 1.92, 95%CI 1.25, 2.96 vs. well/moderately differentiated) but not surgery (Table 2). In a sub-analysis of subjects (n=466) who were treated with either radiation alone or surgery±RT, age was the only significant predictor of treatment, with younger patients more likely to receive surgery (results not shown). We performed multivariable linear regression to analyze the effect of treatment on the number of days spent in the hospital in the 12 months following the therapy (Table 3). There were 639 subjects who received either RT and/or surgery and spent a mean of 40.1 days in the hospital, not necessarily consecutively. Those who underwent RT spent an additional 11.1 days in the hospital, while patients who underwent surgery±RT spent an additional 29 days hospitalized. Patients aged 80–84 spent 10.8 days more than 65–69year-olds in the hospital. In a multivariable Cox proportional hazards model, median survival was 10 months for the notreatment and surgery arms, and 4 months in the RT arm (HR 1.39, 95%CI 1.17–1.66 versus no treatment). Patients with poor or undifferentiated tumors (HR 1.26, 95%CI 1.03–1.54) were also more likely to die compared to those with well to moderately differentiated tumors (results not shown).
Discussion In our sample of elderly men with metastatic prostate cancer, 4.8 % developed MESCC; of these, 50 % were treated with RT while 15 % underwent surgery±RT. We found that black men were more likely to develop MESCC than white men, as were younger men and those without a history of osteoporosis or prior ADT use. Younger age and fewer comorbidities were
associated with undergoing treatment for the spinal cord compression with either surgery or RT. Previous studies suggest that black men present with more advanced-stage disease as compared to white men; this is thought to be largely due to poor access to healthcare [9]. While both observational [10] and interventional [11] data suggest that there is no survival difference between black and white men with advanced-stage prostate cancer, it is still possible that disparities in quality of life may result. Thus, our finding that black men with advanced prostate cancer are more likely to develop MESCC further worsens the burden of disease that blacks experience from prostate cancer. We also found that men without a history of osteoporosis were more likely to develop MESCC. This finding seems counter-intuitive as one would expect that pre-existing osteoporosis would increase the risk of MESCC due to either frailty or pathologic fracture of a vertebral body [3]. However, this may reflect diagnostic bias; the diagnosis of osteoporosis requires men to undergo bone mineral density testing. Only 10 % of men who initiate ADT have a baseline bone density test [12]. Those diagnosed with osteoporosis are often offered treatment with bisphosphonates. Among those with prostate cancer, at least one study showed that only about half of patients received bisphosphonates for the prevention of fractures despite good evidence in its favor [13]. Of those not diagnosed, there may be a higher prevalence of untreated osteoporosis. Therefore, our finding may be due in part to diagnostic bias. Overall, two thirds of subjects with MESCC underwent treatment with either RT alone or surgery±RT. These figures are similar to those reported by Loblaw et al., in a study of MESCC among patients with all types of cancer [14]. These relatively high treatment rates in an elderly cohort, half of whom had at least one comorbidity, suggest that the severity of the symptoms of MESCC prompted intervention. Fifty percent of our cohort received RT while only 15 % had surgery with or without RT. Through the lens of the Patchell study, these results suggest a poorer quality of care delivered to those patients who only received RT, however this interpretation should be tempered. The majority of our sample was diagnosed before the publication of the Patchell randomized trial in 2005 [5], reducing the possible impact of that study’s findings on our patient population. Also, while the Patchell study is the only randomized phase III multicenter trial evaluating the question of surgery versus RT for MESCC, the findings have been criticized as it took 10 years to recruit 50 % of the targeted accrual introducing questions of recruitment bias and validity [15]. The same group later performed a secondary analysis from the clinical trial data evaluating the role of age and treatment outcome and found that for both treatment modalities there was no difference in outcome for patients≥65 years of age [16]. Other retrospective analyses, literature reviews, and meta-analyses have been performed to help answer this question. Loblaw et al. updated their 2005
Support Care Cancer Table 2 Predictors of metastatic epidural spinal cord compression, treatment with radiation, treatment with surgery±radiation, among men diagnosed with stage IV prostate cancer in SEERMedicare, 1991–2007
Category
Predictors of MESCC
Predictors of radiation vs. no treatment
Entire cohort (n=14,800) Odds ratio with 95 % C.I.’s
MESCC cohort (n=711) Odds ratio with 95 % C.I.’s
Predictors of surgery±radiation vs. no treatment MESCC cohort (n=711) Odds ratio with 95 % C.I.’s
Age 65–69
Referent
70–74
1.01 (0.82, 1.24)
0.76 (0.75, 1.22)
0.52 (0.28, 0.96)
75–79
0.79 (0.63, 0.99)
0.88 (0.52, 1.50)
0.68 (0.34, 1.33)
80–84
0.63 (0.49, 0.82)
0.52 (0.29, 0.92)
0.23 (0.10, 0.55)
85+
0.46 (0.34, 0.62)
0.41 (0.22, 0.79)
0.06 (0.01, 0.27)
Race White
Referent
Black
1.75 (1.39, 2.19)
0.96 (0.59, 1.58)
0.77 (0.38, 1.56)
Hispanic
1.30 (0.75, 2.28)
1.67 (0.46, 6.09)
1.20 (0.19, 7.40)
Other
1.80 (1.37, 2.36)
1.53 (0.83, 2.80)
1.31 (0.55, 3.13)
Married
0.86 (0.72, 1.01)
0.92 (0.64, 1.34)
1.02 (0.60, 1.73)
Single/divorced
Referent
Unknown
0.45 (0.28, 0.73)
0.85 (0.28, 2.59)
0.97 (0.21, 4.42)
0.97 (0.50, 1.90)
0.61 (0.27, 1.40)
1.32 (0.59, 2.90)
Marital status
Residence Rural
Referent
Urban
1.24 (0.93, 1.66)
Socioeconomic status First quintile
Referent
Second quintile
0.73 (0.60, 0.94)
0.89 (0.50, 1.59)
Third quintile
0.91 (0.71, 1.18)
0.83 (0.48, 1.46)
0.92 (0.41, 2.03)
Fourth quintile
1.15 (0.89, 1.48)
0.89 (0.50, 1.58)
0.99 (0.44, 2.22)
Fifth quintile
1.02 (0.89, 1.34)
0.95 (0.51, 1.77)
0.86 (0.36, 2.09) Referenta
Tumor grade Well differentiated
Referent
Referenta
Moderately differentiated
1.67 (0.73, 3.82)
Referent
Poorly differentiated
2.28 (0.99, 5.16)
Undifferentiated
3.01 (1.14, 7.94)
Unknown
3.88 (1.69, 8.89)
Referent b
1.92 (1.25, 2.96)
1.37 (0.75, 2.51)b
2.48 (1.49, 4.12)
2.30 (1.15, 4.62)
Comorbidities
Bold entries are statistically significant a
The well and moderately differentiated cases have been merged as the Referent group in these two columns
b
The poorly and undifferentiated groups have been merged in these analyses
None
Referent
Referent
Referent
One
0.91 (0.75, 1.09)
0.55 (0.37, 0.83)
0.59 (0.33, 1.04)
Two+
0.84 (0.68, 1.03)
0.53 (0.34, 0.82)
0.40 (0.20, 0.78)
0.99 (0.97, 1.00)
1.00 (0.97, 1.04)
1.04 (0.99, 1.09)
Year of cancer diagnosis Prior ADT use No
Referent
Yes
0.73 (0.62, 0.86)
Prior osteoporosis No
Referent
Yes
0.63 (0.47, 0.83)
systematic review and guidelines [17]. In it, they conclude that surgery should be considered for patients who are surgical candidates with a good prognosis, while RT should be offered to those who are non-operable, with single doses of RT given to those with poor prognosis and higher doses of RT given to
those with a good prognosis. A Cochrane Review concluded that ambulatory patients with a stable spine might be treated with RT. It reported some evidence of benefit for surgery in ambulatory patient with poor prognostic factors or those who are non-ambulatory, with a short period or paraplegia, and a
Support Care Cancer Table 3 Predictors of hospital length of stay (days) (n=639) for men with stage IV prostate cancer who developed metastatic epidural spinal cord compression, SEER-Medicare, 1991–2007
Treatment None Radiation Surgery±Radiation Patient characteristics Age 65–69 70–74 75–79 80–84 85+ Race White Black Hispanic Other Marital status Married Single/divorced Unknown Residence Rural Urban Socioeconomic status First quintile Second quintile Third quintile
Parameter estimate
95 % C.I.
1.00 11.10 28.57
Referent (4.21, 18.00) (19.23, 37.91)
1.00 −1.38 −0.15
Referent (−9.50, 6.75) (−8.97, 8.67)
10.76 0.29
(0.58, 20.95) (−11.94, 12.52)
1.00 7.47 −9.30 17.21
Referent (−1.39, 16.32) (−30.08, 11.49) (6.62, 27.79)
−11.08 1.00 −13.63
(−17.64, −4.53) Referent (−32.04, 4.78)
1.00 3.84
Referent (−7.79, 15.48)
1.00 −2.56 −5.38
Referent (−12.61, 7.48) (−15.25, 4.49)
Fourth quintile −4.28 (−14.43, 5.87) Fifth quintile −3.01 (−13.91, 7.88) Missing Tumor grade Well differentiated 1.00 Referent Moderately differentiated Poorly differentiated −2.90 (−10.72, 4.91) Undifferentiated 2.81 (−6.20, 11.82) Unknown grade Comorbidities None 1.00 Referent One 1.66 (−5.63, 8.95) Two+ 7.26 (−0.77, 15.28) Year of spinal cord compression diagnosis 0.40 (−0.20, 1.00) Months of follow-up from spinal cord compression diagnosis 1.36 (0.71, 2.02) Bold entries are statistically significant
single lesion [4]. Given the recommendations, it is understandable that the strongest predictors of undergoing either therapy in our study were younger age and fewer comorbidities. It is reassuring that race was not a predictor for receipt of therapy. We found that patients who underwent RT or surgery spent an additional 11 and 29 days, respectively, in the hospital. This compares to 10 days for either group in the Patchell study [5]. While our study has a similar length of stay for the RT groups, patients who underwent surgery in our cohort had longer hospital stays. The difference in length of stay may reflect the difference in patient populations being studied. The Patchell study [5] included younger patients with a variety of cancers leading to MESCC (the median age in the Patchell study was 60 years); therefore, the benefit of surgery and RT for MESCC may be less applicable to older patients with advanced prostate cancer. Our hospital length of stay analysis also suggests that these indications were well chosen by the treating physicians as older men 80–84 years old who underwent surgery and/or radiation spent an extra 11 days in the hospital, as compared to 65–69-year-old patients; and there was a trend toward longer length of stay for those with 2 or more comorbidities who had surgery±RT. Since surgery and RT are palliative therapies, patients should be informed of the realistic goals of these treatments and that symptom relief may be obtained at the cost of spending more time in the hospital, a not insignificant end-of-life consideration. However, left untreated, patients with MESCC will invariably progress to paralysis, incontinence, and shorter survival [18]. It should also be noted that both surgical and radiation techniques have advanced since the publication of the Patchell study, with newer fusion techniques and single fraction radiosurgery allowing more patients to be eligible for therapeutic interventions. A Cox proportional hazard model found that patients who received RT as treatment for MESCC were more likely to die compared to patients who received no treatment. These findings are likely due to a selection bias among patients who received RT rather than no treatment or surgery. Patients who received RT were more likely to have poorly differentiated disease (Table 2), and thus had more aggressive disease and a higher risk of dying. Interpreting these data should be done with caution. Skeletal-related events from prostate cancer are costly, averaging $951 per episode [19]. In an accompanying costeffectiveness analysis to the Patchell randomized trial [18], Thomas found that surgery and radiation were cost-effective, with a baseline incremental cost-effectiveness ratio of $48 per additional day of ambulation and $24,752 per life-year gained (in 2003 US dollars). There are several limitations to our analysis. Details regarding the efficacy of therapy such as ability to ambulate before and after treatment are not available in the SEER-
Support Care Cancer
Medicare database and thus only patterns of care can be examined. Our sample is also limited to men 65 years and older; therefore, our findings may not be generalizable to younger men, although the majority of cases of advanced prostate cancer are diagnosed after age 65. The CPT codes for radiation therapy are not specific to a body part and do not provide data on number of RT fractions given. Therefore, we assumed that patients who received radiation therapy within 60 days following diagnosis of MESCC were receiving treatment for the MESCC. As described above, we cannot determine if patients received oral bisphosphonates for the prevention of osteoporosis, which might have influenced the impact of osteoporosis on the development of MESCC. In this large population-based study, we have demonstrated that black men with advanced prostate cancer are more likely than white men to develop MESCC. Older patients with spinal cord compression and those with comorbidities are less likely to be treated with either form of treatment in this palliative care setting. Acknowledgments This study was funded in part by a grant from the Department of Defense (PC094372). Conflict of interest None of the authors has a conflict of interest to report. No author has a financial relationship with the organization that sponsored the research. The authors have full control of the data used for analysis in this study. We agree to allow the Journal to review the data, if requested.
References 1. Siegel R, Naishadham D, Jemal A (2012) Cancer statistics. CA Cancer J Clin 62(1):10–29 2. Bubendorf L, Schopfer A, Wagner U et al (2000) Metastatic patterns of prostate cancer: an autopsy study of 1,589 patients. Hum Pathol 31(5):578–583 3. Lee RJ, Saylor PJ, Smith MR (2011) Treatment and prevention of bone complications from prostate cancer. Bone 48(1):88–95 4. George R, Jeba J, Ramkumar G, Chacko AG, Leng M, Tharyan P (2008) Interventions for the treatment of metastatic extradural spinal cord compression in adults. The Cochrane Database of Systematic Reviews. (4):CD006716 5. Patchell RA, Tibbs PA, Regine WF et al (2005) Direct decompressive surgical resection in the treatment of spinal cord
compression caused by metastatic cancer: a randomised trial. Lancet 366(9486):643–648 6. Du XL, Fang S, Vernon SW et al (2007) Racial disparities and socioeconomic status in association with survival in a large population-based cohort of elderly patients with colon cancer. Cancer 110(3):660–669 7. Charlson ME, Pompei P, Ales KL, MacKenzie CR (1987) A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J chronic dis 40(5): 373–383 8. Klabunde CN, Potosky AL, Legler JM, Warren JL (2000) Development of a comorbidity index using physician claims data. J Clin Epidemiol 53(12):1258–1267 9. Hoffman RM, Gilliland FD, Eley JW et al (2001) Racial and ethnic differences in advanced-stage prostate cancer: the Prostate Cancer Outcomes Study. J Natl Cancer Inst 93(5): 388–395 10. Polednak AP (2003) Black-white differences in survival from latestage prostate cancer. Ethn Dis 13(2):220–225 11. Tangen CM, Hussain MH, Higano CS et al (2012) Improved overall survival trends of men with newly diagnosed M1 prostate cancer: a SWOG phase III trial experience (S8494, S8894 and S9346). The Journal of urology 188(4):1164–1169 12. Morgans AK, Smith MR, O’Malley AJ, Keating NL (2013) Bone density testing among prostate cancer survivors treated with androgen-deprivation therapy. Cancer 119(4):863–870 13. Freedland SJ, Richhariya A, Wang H, Chung K, Shore ND (2012) Treatment patterns in patients with prostate cancer and bone metastasis among US community-based urology group practices. Urology 80(2):293–298 14. Loblaw DA, Laperriere NJ, Mackillop WJ (2003) A populationbased study of malignant spinal cord compression in Ontario. Clin Oncol 15(4):211–217 15. van den Bent MJ (2005) Surgical resection improves outcome in metastatic epidural spinal cord compression. Lancet 366(9486):609– 610 16. Chi JH, Gokaslan Z, McCormick P, Tibbs PA, Kryscio RJ, Patchell RA (2009) Selecting treatment for patients with malignant epidural spinal cord compression-does age matter? Results from a randomized clinical trial. Spine 34(5):431–435 17. Loblaw DA, Mitera G, Ford M, Laperriere NJ (2012) A 2011 updated systematic review and clinical practice guideline for the management of malignant extradural spinal cord compression. Int J Radiat Oncol Biol Physics 84(2):312–317 18. Thomas KC, Nosyk B, Fisher CG et al (2006) Cost-effectiveness of surgery plus radiotherapy versus radiotherapy alone for metastatic epidural spinal cord compression. Int J Radiat Oncol Biol Physics 66(4):1212–1218 19. Hagiwara M, Delea TE, Saville MW, Chung K ( 2012) Healthcare utilization and costs associated with skeletal-related events in prostate cancer patients with bone metastases. Prostate Cancer Prostatic Dis. Nov 13
ORIGINAL ARTICLE
Metastatic epidural spinal cord compression among elderly patients with advanced prostate cancer Benjamin A. Spencer & Jin Joo Shim & Dawn L. Hershman & Brad E. Zacharia & Emerson A. Lim & Mitchell C. Benson & Alfred I. Neugut
Received: 25 March 2013 / Accepted: 25 December 2013 # Springer-Verlag Berlin Heidelberg 2014
Abstract Background A recent randomized trial demonstrated that for metastatic epidural spinal cord compression (MESCC), a complication of advanced prostate cancer, surgical decompression may be more effective than external beam radiation therapy (RT). We investigated predictors of MESCC, its treatment, and its impact on hospital length of stay for patients with advanced prostate cancer. Methods We used the SEER-Medicare database to identify patients >65 years with stage IV (n=14,800) prostate cancer. We used polytomous logistic regression to compare those with
The linked SEER-Medicare database was used in this study. The interpretation and reporting of these data are the sole responsibility of the authors. B. A. Spencer : M. C. Benson Department of Urology, College of Physicians and Surgeons, Columbia University, New York, NY, USA B. A. Spencer : J. J. Shim : D. L. Hershman : A. I. Neugut Department of Epidemiology, Mailman School of Public Health, Columbia University, New York, NY, USA D. L. Hershman : E. A. Lim : A. I. Neugut Department of Medicine, College of Physicians and Surgeons, Columbia University, New York, NY, USA B. E. Zacharia Department of Neurological Surgery, College of Physicians and Surgeons, Columbia University, New York, NY, USA B. A. Spencer : D. L. Hershman : E. A. Lim : M. C. Benson : A. I. Neugut Herbert Irving Comprehensive Cancer Center, College of Physicians and Surgeons, Columbia University, New York, NY, USA A. I. Neugut (*) Columbia University Medical Center, 722 West 168th Street, Room 725, New York, NY 10032, USA e-mail: [email protected]
and without MESCC and those hospitalized for treatment with surgical decompression and/or RT. Results MESCC developed in 711 (5 %) of patients, among whom 359 (50 %) received RT and 107 (15 %) underwent surgery±RT. Median survival was 10 months. MESCC was more likely among patients who were black (OR 1.75, 95 %CI 1.39–2.19 vs. white) and had high-grade tumors (OR 3.01, 95 %CI 1.14–7.94), and less likely in those younger; with prior hormonal therapy (OR 0.73, 95 %CI 0.62–0.86); or with osteoporosis (OR 0.63, 95 %CI 0.47–0.83). Older patients were less likely to undergo either RT or surgery, as were those with ≥1 comorbidity. Patients with high-grade tumors were more likely to undergo RT (OR 1.92, 95 %CI 1.25–2.96). Those who underwent RT or surgery spent an additional 11 and 29 days, respectively, hospitalized. Conclusions We found that black men with metastatic prostate cancer are more likely to develop MESCC than whites. RT was more commonly utilized for treatment than surgery, but the elderly and those with comorbidities were unlikely to receive either treatment. Keywords Prostate cancer . Metastatic epidural spinal cord compression . Palliative care . SEER-Medicare
Introduction Each year, about 4 % of men diagnosed with prostate cancer are diagnosed with distant or metastatic disease [1]. The most common site for metastatic spread in prostate cancer is bone [2]. Bone metastases can cause pain, fractures, and metastatic epidural spinal cord compression (MESCC) [3]. The resulting edema, venous congestion, and demyelination can lead to irreversible spinal cord infarction if not treated promptly. With life expectancy for patients with MESCC estimated at 4 months, the decision to relieve the compression must be
Support Care Cancer
made with realistic goals [4]. However, the potential improvement in functional status (pain, ambulation, and urinary continence) and quality of life can be substantial. Relief of spinal cord compression can be achieved either with direct surgical decompression, external beam radiation therapy (RT) or both. In 2005, a randomized trial by Patchell et al. demonstrated a benefit from circumferential decompression, as compared to posterior laminectomy, followed by postoperative RT in patients with metastatic cancer presenting with signs and symptoms of cord compression and a true displacement of the spinal cord by an epidural mass [5]. For subjects who were randomized to both surgery and RT, 84 % were able to walk after treatment as compared to 57 % in the RT alone arm (p=0.001) and this ambulation persisted for 122 days as compared to 13 days in the RT alone group (p=0.003). Though this study is the only randomized trial evaluating the use of RT versus surgery for MESCC, it has been criticized because of slow patient accrual, suggesting a recruitment bias. The objective of our study was to identify predictors of MESCC, determine the patterns of use for surgery and RT, and investigate whether these interventions decrease the number of days spent in the hospital. We were also interested in whether disparities due to race or access to care influence the development of MESCC and its subsequent treatment.
whose only reporting source was autopsy or death certificate, whose reason for entitlement was not age, or whose date of death differed between SEER and Medicare by more than 3 months were also excluded. Among the remaining 139,627 patients, we selected a cohort that was diagnosed with American Joint Cancer Committee (AJCC) stage IV prostate cancer and survived at least 90 days following diagnosis. We then identified patients who had at least one billing claim with a diagnosis of spinal cord compression (ICD-9 Code 336.9) no more than 30 days prior to the prostate cancer diagnosis. This would allow for the possibility that the cord compression was the presenting symptom for the prostate cancer. Demographic variables SEER-Medicare provides the age, race/ethnicity, marital status, and tumor grade as categorical variables. Age categories are ordinal with 5-year increments between categories. Race/ethnicity was described as white, black, Hispanic, and Other/Missing. Marital status at time of cancer diagnosis was categorized as Married, Single/Divorced/Separated/Widowed, or Unknown. Tumor grade was categorized as well, moderately, poorly, and undifferentiated. Socioeconomic status
Methods Data source We analyzed data from the Surveillance, Epidemiology, and End Results (SEER)—Medicare database. The SEER database includes information on cancer diagnoses, tumor histology, stage of disease, treatments, socioeconomic status at the census tract and zip code level, survival and demographic characteristics. The Medicare database provides information on Medicare A (inpatient) and B (outpatient) eligibility, reason for Medicare entitlement, and diagnoses. The physician and hospital claim files, as well as inpatient claim files, were used to search for the claims on diagnoses or surgical procedures. The SEER and Medicare databases were combined by linking these patients by their unique patient identification number. Cohort selection We identified individuals who were 65 years or older at the time of cancer diagnosis and received a pathologically confirmed primary diagnosis of prostate cancer (SEER Site Code 54) as their first cancer between January 1, 1991, and December 31, 2007. Patients who were enrolled in a non-Medicare health maintenance organization or not covered by Medicare Part A and B over the same period were excluded. Patients
We followed the guidelines by Du et al. [6] to create a socioeconomic status score by equally weighting income, education, and poverty level provided from the 2000 census tract data. Patients were assigned a rank score from 0 to 4, with 0 being the lowest score. Approximately 1.6 % (N=233) of the cohort lacked sufficient data in one or more categories and were excluded from the analyses. Assessment of comorbid disease Using the comorbid conditions identified by Charlson et al. [7], we searched for diagnostic codes of the ICD-9 Clinical Modification and Healthcare Common Procedure Coding System (HCPCS) for 18 medical conditions. Each medical condition was assigned a weight and subsequently used to calculate the final index score. In order to obtain the complete diagnosis and medical claims, the physician and outpatient claims, as well as hospital claims, were included in the comorbidity calculation, as described by Klabunde et al. [8]. Claims submitted from 13 months prior to 4 months following the date of cancer diagnosis were considered in the comorbidity index calculation. Treatment characteristics Receipt of radiation and surgery were extracted from the Medicare database by searching the HCPCS, Current
Support Care Cancer
Procedural Terminology (CPT) codes, ICD-9 Clinical Modification diagnostic codes and procedure codes from the national claims history files, outpatient statistical analysis files, and Medicare provider analysis and review files. We excluded radiation and surgery claims that were billed more than 15 days prior or 60 days following the date of diagnosis of spinal cord compression because patients could have received these treatments for medical conditions other than spinal cord compression. The CPT codes 63001, 63003, 63005, 63012, 63015–63017, 63020, 63030, 63035, 63040, 63042–63048, 63050–63051, 63055–63057, 63064, 63066, 63075–63076, 63078, 63081–63082, 63065–83088, 63091, 63101–63013, 63275–63278, 63280–63283, 63286–63287, 63290, and 63295 and ICD-9 CM codes 03.0, 03.09, 80.5, 80.51, 80.52, 80.59, 81.0 81.00 81.04, 81.05, and 81.09 corresponded to surgical decompression. The ICD-9 CM codes V58.0, V66.1, V67.1, 92.21–92.26, 92.29–92.33, and 92.39 and CPT/ HCPCS codes 77401–77409, 77411–77414, 77416–77421, 77427, 77431–77432, 77435, 77470, and 77499 were used to identify patients who received radiation therapy. Based on the claims, patients were assigned to one of three treatment groups—radiation alone, surgery alone or in conjunction with radiation (+/- RT), and neither. Receipt of androgen deprivation therapy (ADT) with a GnRH agonist (HCPCS codes J1950, J3315, J9202, J9217-J9219, J9225) was assigned for patients who had at least one claim for receipt of ADT prior to the MESCC. Statistical analysis We used the chi-square test to compare the distributions of those who developed spinal cord compression with those who did not (see Table 1), and multivariable logistic regression to identify predictors of spinal cord compression, adjusting for the measured clinical and demographic characteristics of the patients. Polytomous logistic regression models were used to test associations between patient demographics and treatment assignment. All statistical tests were two-sided with an alpha of 0.05. Statistical analyses were performed using SAS version 9.2 (Cary, NC). We obtained approval for this study from the Columbia University Medical Center Institutional Review Board.
Results We identified 14,800 men with stage IV prostate cancer who met our eligibility criteria. Table 1 shows their demographic and clinical characteristics. The majority of the stage IV patients were white (79 %), married (67 %), lived in an urban area (89 %), had high-grade tumors (53 %), had no comorbidities (52 %), had used ADT prior to the diagnosis of
Table 1 Socio-demographic and clinical characteristics of men >65 years with stage IV prostate cancer in SEER-Medicare, 1991–2007 Characteristic
Age 65–69 70–74 75–79 80–84 85+ Race White Black Hispanic Other Marital status Married
Entire sample
Spinal cord compression sample
(n=14,800) n (%) 3,065 (20.7 3,925 (26.5 3,210 (21.7 2,568 (17.4 2,032 (13.7
(n=711) n (%) 170 (23.9 %) 225 (31.7 %) 149 (21.0 %) 102 (14.4 %) 65 (9.1 %)
%) %) %) %) %)
11,745 (79.4 %) 1,854 (12.5 %) 273 (1.8 %) 928 (6.3 %)
488 (68.6 %) 143 (20.1 %) 14 (2.0 %) 66 (9.3 %)
9,870 (66.7 %)
458 (64.4 %)
Single/divorced/widowed Unknown Residence Urban Rural Socioeconomic status First quintile Second quintile Third quintile Fourth quintile Fifth quintile Missing Tumor grade Well/moderately differentiated Poorly differentiated Undifferentiated Unknown Comorbidities
4,220 (28.5 %) 710 (4.8 %)
235 (33.1 %) 18 (2.5 %)
13,135 (88.8 %) 1,665 (11.3 %)
654 (92.0 %) 57 (8.0 %)
2,610 (17.6 3,223 (21.8 3,325 (22.5 2,936 (19.8 2,706 (18.3
158 (22.2 122 (17.2 146 (20.5 158 (22.2 127 (17.9
4,152 (28.0 %) 7,591 (51.3 %) 242 (1.6 %) 2,815 (19.0 %)
153 (21.5 %) 351 (49.4 %) 15 (2.1 %) 192 (27.0 %)
None One ≥Two Prior ADT use No Yes Prior osteoporosis No Yes Treatment None Radiation Surgery
7,696 (52.0 %) 3,871 (26.2 %) 3,233 (21.8 %)
399 (56.1 %) 176 (24.8 %) 136 (19.1 %)
5,815 (39.3 %) 8,985 (60.7 %)
370 (52.0 %) 341 (48.0 %)
12,828 (86.7 %) 1,972 (13.3 %)
57 (8.0 %) 654 (92.0 %)
NA NA NA
245 (34 %) 359 (50 %) 107 (15 %)
%) %) %) %) %)
%) %) %) %) %)
Support Care Cancer
MESCC (61 %), and had no prior history of osteoporosis (87 %). Almost 5 % (n=711) developed MESCC, among whom 359 (50 %) underwent palliative radiation therapy and 107 (15 %) underwent surgical decompression±RT. In the multivariable analysis, we found that MESCC was more likely to occur among those who were younger; black (OR 1.75, 95%CI 1.39, 2.19 vs. white); had undifferentiated tumors (OR 3.01, 95%CI 1.14, 7.94 vs. well differentiated); were diagnosed earlier in the cohort; did not use ADT prior to the diagnosis of MESCC (OR 0.73, 95%CI 0.62, 0.86); and had no prior history of osteoporosis (OR 0.63, 95%CI 0.47, 0.83) (Table 2). There were 711 stage IV patients (4.8 %) who developed MESCC. Of these, 359 (50 %) underwent RT, while 107 (15 %) underwent spinal surgery. Among subjects who developed MESCC, older patients were less likely to undergo either radiation or surgery. Similarly, those with comorbidities were also less likely to undergo either radiation (2+ comorbidities: OR 0.53, 95%CI 0.34, 0.82 vs no comorbidities) or surgery± radiation (2+ comorbidities: OR 0.40, 95%CI 0.20, 0.78 vs no comorbidities). Patients with high-grade tumors were more likely to undergo RT (poorly/undifferentiated: OR 1.92, 95%CI 1.25, 2.96 vs. well/moderately differentiated) but not surgery (Table 2). In a sub-analysis of subjects (n=466) who were treated with either radiation alone or surgery±RT, age was the only significant predictor of treatment, with younger patients more likely to receive surgery (results not shown). We performed multivariable linear regression to analyze the effect of treatment on the number of days spent in the hospital in the 12 months following the therapy (Table 3). There were 639 subjects who received either RT and/or surgery and spent a mean of 40.1 days in the hospital, not necessarily consecutively. Those who underwent RT spent an additional 11.1 days in the hospital, while patients who underwent surgery±RT spent an additional 29 days hospitalized. Patients aged 80–84 spent 10.8 days more than 65–69year-olds in the hospital. In a multivariable Cox proportional hazards model, median survival was 10 months for the notreatment and surgery arms, and 4 months in the RT arm (HR 1.39, 95%CI 1.17–1.66 versus no treatment). Patients with poor or undifferentiated tumors (HR 1.26, 95%CI 1.03–1.54) were also more likely to die compared to those with well to moderately differentiated tumors (results not shown).
Discussion In our sample of elderly men with metastatic prostate cancer, 4.8 % developed MESCC; of these, 50 % were treated with RT while 15 % underwent surgery±RT. We found that black men were more likely to develop MESCC than white men, as were younger men and those without a history of osteoporosis or prior ADT use. Younger age and fewer comorbidities were
associated with undergoing treatment for the spinal cord compression with either surgery or RT. Previous studies suggest that black men present with more advanced-stage disease as compared to white men; this is thought to be largely due to poor access to healthcare [9]. While both observational [10] and interventional [11] data suggest that there is no survival difference between black and white men with advanced-stage prostate cancer, it is still possible that disparities in quality of life may result. Thus, our finding that black men with advanced prostate cancer are more likely to develop MESCC further worsens the burden of disease that blacks experience from prostate cancer. We also found that men without a history of osteoporosis were more likely to develop MESCC. This finding seems counter-intuitive as one would expect that pre-existing osteoporosis would increase the risk of MESCC due to either frailty or pathologic fracture of a vertebral body [3]. However, this may reflect diagnostic bias; the diagnosis of osteoporosis requires men to undergo bone mineral density testing. Only 10 % of men who initiate ADT have a baseline bone density test [12]. Those diagnosed with osteoporosis are often offered treatment with bisphosphonates. Among those with prostate cancer, at least one study showed that only about half of patients received bisphosphonates for the prevention of fractures despite good evidence in its favor [13]. Of those not diagnosed, there may be a higher prevalence of untreated osteoporosis. Therefore, our finding may be due in part to diagnostic bias. Overall, two thirds of subjects with MESCC underwent treatment with either RT alone or surgery±RT. These figures are similar to those reported by Loblaw et al., in a study of MESCC among patients with all types of cancer [14]. These relatively high treatment rates in an elderly cohort, half of whom had at least one comorbidity, suggest that the severity of the symptoms of MESCC prompted intervention. Fifty percent of our cohort received RT while only 15 % had surgery with or without RT. Through the lens of the Patchell study, these results suggest a poorer quality of care delivered to those patients who only received RT, however this interpretation should be tempered. The majority of our sample was diagnosed before the publication of the Patchell randomized trial in 2005 [5], reducing the possible impact of that study’s findings on our patient population. Also, while the Patchell study is the only randomized phase III multicenter trial evaluating the question of surgery versus RT for MESCC, the findings have been criticized as it took 10 years to recruit 50 % of the targeted accrual introducing questions of recruitment bias and validity [15]. The same group later performed a secondary analysis from the clinical trial data evaluating the role of age and treatment outcome and found that for both treatment modalities there was no difference in outcome for patients≥65 years of age [16]. Other retrospective analyses, literature reviews, and meta-analyses have been performed to help answer this question. Loblaw et al. updated their 2005
Support Care Cancer Table 2 Predictors of metastatic epidural spinal cord compression, treatment with radiation, treatment with surgery±radiation, among men diagnosed with stage IV prostate cancer in SEERMedicare, 1991–2007
Category
Predictors of MESCC
Predictors of radiation vs. no treatment
Entire cohort (n=14,800) Odds ratio with 95 % C.I.’s
MESCC cohort (n=711) Odds ratio with 95 % C.I.’s
Predictors of surgery±radiation vs. no treatment MESCC cohort (n=711) Odds ratio with 95 % C.I.’s
Age 65–69
Referent
70–74
1.01 (0.82, 1.24)
0.76 (0.75, 1.22)
0.52 (0.28, 0.96)
75–79
0.79 (0.63, 0.99)
0.88 (0.52, 1.50)
0.68 (0.34, 1.33)
80–84
0.63 (0.49, 0.82)
0.52 (0.29, 0.92)
0.23 (0.10, 0.55)
85+
0.46 (0.34, 0.62)
0.41 (0.22, 0.79)
0.06 (0.01, 0.27)
Race White
Referent
Black
1.75 (1.39, 2.19)
0.96 (0.59, 1.58)
0.77 (0.38, 1.56)
Hispanic
1.30 (0.75, 2.28)
1.67 (0.46, 6.09)
1.20 (0.19, 7.40)
Other
1.80 (1.37, 2.36)
1.53 (0.83, 2.80)
1.31 (0.55, 3.13)
Married
0.86 (0.72, 1.01)
0.92 (0.64, 1.34)
1.02 (0.60, 1.73)
Single/divorced
Referent
Unknown
0.45 (0.28, 0.73)
0.85 (0.28, 2.59)
0.97 (0.21, 4.42)
0.97 (0.50, 1.90)
0.61 (0.27, 1.40)
1.32 (0.59, 2.90)
Marital status
Residence Rural
Referent
Urban
1.24 (0.93, 1.66)
Socioeconomic status First quintile
Referent
Second quintile
0.73 (0.60, 0.94)
0.89 (0.50, 1.59)
Third quintile
0.91 (0.71, 1.18)
0.83 (0.48, 1.46)
0.92 (0.41, 2.03)
Fourth quintile
1.15 (0.89, 1.48)
0.89 (0.50, 1.58)
0.99 (0.44, 2.22)
Fifth quintile
1.02 (0.89, 1.34)
0.95 (0.51, 1.77)
0.86 (0.36, 2.09) Referenta
Tumor grade Well differentiated
Referent
Referenta
Moderately differentiated
1.67 (0.73, 3.82)
Referent
Poorly differentiated
2.28 (0.99, 5.16)
Undifferentiated
3.01 (1.14, 7.94)
Unknown
3.88 (1.69, 8.89)
Referent b
1.92 (1.25, 2.96)
1.37 (0.75, 2.51)b
2.48 (1.49, 4.12)
2.30 (1.15, 4.62)
Comorbidities
Bold entries are statistically significant a
The well and moderately differentiated cases have been merged as the Referent group in these two columns
b
The poorly and undifferentiated groups have been merged in these analyses
None
Referent
Referent
Referent
One
0.91 (0.75, 1.09)
0.55 (0.37, 0.83)
0.59 (0.33, 1.04)
Two+
0.84 (0.68, 1.03)
0.53 (0.34, 0.82)
0.40 (0.20, 0.78)
0.99 (0.97, 1.00)
1.00 (0.97, 1.04)
1.04 (0.99, 1.09)
Year of cancer diagnosis Prior ADT use No
Referent
Yes
0.73 (0.62, 0.86)
Prior osteoporosis No
Referent
Yes
0.63 (0.47, 0.83)
systematic review and guidelines [17]. In it, they conclude that surgery should be considered for patients who are surgical candidates with a good prognosis, while RT should be offered to those who are non-operable, with single doses of RT given to those with poor prognosis and higher doses of RT given to
those with a good prognosis. A Cochrane Review concluded that ambulatory patients with a stable spine might be treated with RT. It reported some evidence of benefit for surgery in ambulatory patient with poor prognostic factors or those who are non-ambulatory, with a short period or paraplegia, and a
Support Care Cancer Table 3 Predictors of hospital length of stay (days) (n=639) for men with stage IV prostate cancer who developed metastatic epidural spinal cord compression, SEER-Medicare, 1991–2007
Treatment None Radiation Surgery±Radiation Patient characteristics Age 65–69 70–74 75–79 80–84 85+ Race White Black Hispanic Other Marital status Married Single/divorced Unknown Residence Rural Urban Socioeconomic status First quintile Second quintile Third quintile
Parameter estimate
95 % C.I.
1.00 11.10 28.57
Referent (4.21, 18.00) (19.23, 37.91)
1.00 −1.38 −0.15
Referent (−9.50, 6.75) (−8.97, 8.67)
10.76 0.29
(0.58, 20.95) (−11.94, 12.52)
1.00 7.47 −9.30 17.21
Referent (−1.39, 16.32) (−30.08, 11.49) (6.62, 27.79)
−11.08 1.00 −13.63
(−17.64, −4.53) Referent (−32.04, 4.78)
1.00 3.84
Referent (−7.79, 15.48)
1.00 −2.56 −5.38
Referent (−12.61, 7.48) (−15.25, 4.49)
Fourth quintile −4.28 (−14.43, 5.87) Fifth quintile −3.01 (−13.91, 7.88) Missing Tumor grade Well differentiated 1.00 Referent Moderately differentiated Poorly differentiated −2.90 (−10.72, 4.91) Undifferentiated 2.81 (−6.20, 11.82) Unknown grade Comorbidities None 1.00 Referent One 1.66 (−5.63, 8.95) Two+ 7.26 (−0.77, 15.28) Year of spinal cord compression diagnosis 0.40 (−0.20, 1.00) Months of follow-up from spinal cord compression diagnosis 1.36 (0.71, 2.02) Bold entries are statistically significant
single lesion [4]. Given the recommendations, it is understandable that the strongest predictors of undergoing either therapy in our study were younger age and fewer comorbidities. It is reassuring that race was not a predictor for receipt of therapy. We found that patients who underwent RT or surgery spent an additional 11 and 29 days, respectively, in the hospital. This compares to 10 days for either group in the Patchell study [5]. While our study has a similar length of stay for the RT groups, patients who underwent surgery in our cohort had longer hospital stays. The difference in length of stay may reflect the difference in patient populations being studied. The Patchell study [5] included younger patients with a variety of cancers leading to MESCC (the median age in the Patchell study was 60 years); therefore, the benefit of surgery and RT for MESCC may be less applicable to older patients with advanced prostate cancer. Our hospital length of stay analysis also suggests that these indications were well chosen by the treating physicians as older men 80–84 years old who underwent surgery and/or radiation spent an extra 11 days in the hospital, as compared to 65–69-year-old patients; and there was a trend toward longer length of stay for those with 2 or more comorbidities who had surgery±RT. Since surgery and RT are palliative therapies, patients should be informed of the realistic goals of these treatments and that symptom relief may be obtained at the cost of spending more time in the hospital, a not insignificant end-of-life consideration. However, left untreated, patients with MESCC will invariably progress to paralysis, incontinence, and shorter survival [18]. It should also be noted that both surgical and radiation techniques have advanced since the publication of the Patchell study, with newer fusion techniques and single fraction radiosurgery allowing more patients to be eligible for therapeutic interventions. A Cox proportional hazard model found that patients who received RT as treatment for MESCC were more likely to die compared to patients who received no treatment. These findings are likely due to a selection bias among patients who received RT rather than no treatment or surgery. Patients who received RT were more likely to have poorly differentiated disease (Table 2), and thus had more aggressive disease and a higher risk of dying. Interpreting these data should be done with caution. Skeletal-related events from prostate cancer are costly, averaging $951 per episode [19]. In an accompanying costeffectiveness analysis to the Patchell randomized trial [18], Thomas found that surgery and radiation were cost-effective, with a baseline incremental cost-effectiveness ratio of $48 per additional day of ambulation and $24,752 per life-year gained (in 2003 US dollars). There are several limitations to our analysis. Details regarding the efficacy of therapy such as ability to ambulate before and after treatment are not available in the SEER-
Support Care Cancer
Medicare database and thus only patterns of care can be examined. Our sample is also limited to men 65 years and older; therefore, our findings may not be generalizable to younger men, although the majority of cases of advanced prostate cancer are diagnosed after age 65. The CPT codes for radiation therapy are not specific to a body part and do not provide data on number of RT fractions given. Therefore, we assumed that patients who received radiation therapy within 60 days following diagnosis of MESCC were receiving treatment for the MESCC. As described above, we cannot determine if patients received oral bisphosphonates for the prevention of osteoporosis, which might have influenced the impact of osteoporosis on the development of MESCC. In this large population-based study, we have demonstrated that black men with advanced prostate cancer are more likely than white men to develop MESCC. Older patients with spinal cord compression and those with comorbidities are less likely to be treated with either form of treatment in this palliative care setting. Acknowledgments This study was funded in part by a grant from the Department of Defense (PC094372). Conflict of interest None of the authors has a conflict of interest to report. No author has a financial relationship with the organization that sponsored the research. The authors have full control of the data used for analysis in this study. We agree to allow the Journal to review the data, if requested.
References 1. Siegel R, Naishadham D, Jemal A (2012) Cancer statistics. CA Cancer J Clin 62(1):10–29 2. Bubendorf L, Schopfer A, Wagner U et al (2000) Metastatic patterns of prostate cancer: an autopsy study of 1,589 patients. Hum Pathol 31(5):578–583 3. Lee RJ, Saylor PJ, Smith MR (2011) Treatment and prevention of bone complications from prostate cancer. Bone 48(1):88–95 4. George R, Jeba J, Ramkumar G, Chacko AG, Leng M, Tharyan P (2008) Interventions for the treatment of metastatic extradural spinal cord compression in adults. The Cochrane Database of Systematic Reviews. (4):CD006716 5. Patchell RA, Tibbs PA, Regine WF et al (2005) Direct decompressive surgical resection in the treatment of spinal cord
compression caused by metastatic cancer: a randomised trial. Lancet 366(9486):643–648 6. Du XL, Fang S, Vernon SW et al (2007) Racial disparities and socioeconomic status in association with survival in a large population-based cohort of elderly patients with colon cancer. Cancer 110(3):660–669 7. Charlson ME, Pompei P, Ales KL, MacKenzie CR (1987) A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J chronic dis 40(5): 373–383 8. Klabunde CN, Potosky AL, Legler JM, Warren JL (2000) Development of a comorbidity index using physician claims data. J Clin Epidemiol 53(12):1258–1267 9. Hoffman RM, Gilliland FD, Eley JW et al (2001) Racial and ethnic differences in advanced-stage prostate cancer: the Prostate Cancer Outcomes Study. J Natl Cancer Inst 93(5): 388–395 10. Polednak AP (2003) Black-white differences in survival from latestage prostate cancer. Ethn Dis 13(2):220–225 11. Tangen CM, Hussain MH, Higano CS et al (2012) Improved overall survival trends of men with newly diagnosed M1 prostate cancer: a SWOG phase III trial experience (S8494, S8894 and S9346). The Journal of urology 188(4):1164–1169 12. Morgans AK, Smith MR, O’Malley AJ, Keating NL (2013) Bone density testing among prostate cancer survivors treated with androgen-deprivation therapy. Cancer 119(4):863–870 13. Freedland SJ, Richhariya A, Wang H, Chung K, Shore ND (2012) Treatment patterns in patients with prostate cancer and bone metastasis among US community-based urology group practices. Urology 80(2):293–298 14. Loblaw DA, Laperriere NJ, Mackillop WJ (2003) A populationbased study of malignant spinal cord compression in Ontario. Clin Oncol 15(4):211–217 15. van den Bent MJ (2005) Surgical resection improves outcome in metastatic epidural spinal cord compression. Lancet 366(9486):609– 610 16. Chi JH, Gokaslan Z, McCormick P, Tibbs PA, Kryscio RJ, Patchell RA (2009) Selecting treatment for patients with malignant epidural spinal cord compression-does age matter? Results from a randomized clinical trial. Spine 34(5):431–435 17. Loblaw DA, Mitera G, Ford M, Laperriere NJ (2012) A 2011 updated systematic review and clinical practice guideline for the management of malignant extradural spinal cord compression. Int J Radiat Oncol Biol Physics 84(2):312–317 18. Thomas KC, Nosyk B, Fisher CG et al (2006) Cost-effectiveness of surgery plus radiotherapy versus radiotherapy alone for metastatic epidural spinal cord compression. Int J Radiat Oncol Biol Physics 66(4):1212–1218 19. Hagiwara M, Delea TE, Saville MW, Chung K ( 2012) Healthcare utilization and costs associated with skeletal-related events in prostate cancer patients with bone metastases. Prostate Cancer Prostatic Dis. Nov 13
